Can you tell how popular an artist is, just by analyzing a single song? Can you tell in advance how many streams that song will get on Spotify?
In this small-scale pilot study, a collaboration between Musiio and Meddling A&R, we aim to determine whether our Hit Potential Algorithm can be used to predict the success of a song and the size of that song’s artist(s)’ fanbase.
What is Hit Potential?
Our Hit Potential Algorithm assesses the commercial potential of songs based on its audio alone and awards each analyzed track a score from 0-100.
Meddling A&R sent us a list of 2541 recently-released songs, which we analyzed with our Hit Potential Algorithm and sorted by Hit Potential score (HP Score).
We identified two extreme ranges for analysis
- Songs with HP Score above 80 – there were 31 of these. We randomly selected 16 of these for the study.
- Songs with HP Score below 53 – there were ~400 of these. Many of these (instrumental beats for example) were clearly going to perform poorly (since instrumental music rarely makes it to the charts), so we selected 10 for the study that we felt would at least have some chance of succeeding, to create a more reliable test.
Monitoring Success Metrics
These songs were released between June 15 2021 and July 2 2021. We monitored their progress as well as success-related metrics and collected this data on the 30th day of release for each individual song so that each song would have spent and equal amount of time in circulation before its success was measured.
Here are 2 key metrics that we used to measure 2 types of success.
- Song Popularity: Average of YouTube and Spotify Streams
- Artist Popularity: As a metric for an artist’s fanbase/following, we took the average of Spotify Monthly Listens, Facebook Followers, and Instagram followers. If more than one artist was involved on a song, we added up each artist’s metrics and then averaged them out.
Songs with higher HP scores were streamed more
We found that songs with <53 HP Scores far less likely to receive a high number of streams, with ~60% of the songs in the >80 HP Score set receiving higher average streaming numbers.
In addition, the most streamed song in the <53 HP Score set had fewer than 0.5M streams, while the most streamed song in the >83 HP Score set received 8.5M streams after just 30 days.
Here’s the most popular (8.5M streams after 30 days) song from this study:
Songs with higher HP scores came from more popular artists
Songs with higher HP scores tended to come from artists with larger fanbases, but it is clear from the graphs that while a correlation exists between HP score and Artist Popularity, it is not as strong as the correlation between HP score and Song Popularity.
This is expected, since HP Score only takes the audio of a single song into account when awarding a score, and does not take any other contextual information such as branding, culture etc into consideration.
Here’s the song from the most popular artist (19.2M Artist Popularity) in this study.
Initial Thoughts On Useful Categorization
In a sense, there are 4 possible categories that may emerge from such analysis:
- Low HP, Low Streams “Hidden Niche”
- Low HP, High Streams “Niche Hits”
- High HP, High Streams “Hits”
- High HP, Low Streams “Hidden Hits”
At this stage, all recommendations are thought experiments and will require further study of an individual record label’s catalog to truly identify the best strategy for that particular label. For this article we will focus on categories 1 and 4.
Low HP, Low Streams “Hidden Niche”
We know that low Hit Potential scores indicate a low likelihood of commercial success because songs receiving such scores do not sound like the commercially successful hits of today. This may indicate the 2 following (or more) possibilities:
- Certain low HP songs actually sound “bad” – poor production, out-of-tune singing, bad rhythmic interpretation etc. All the usual hallmarks of songs that we can broadly consider “bad”
- Certain low HP songs may not sound “bad”, but they may sound outdated or highly unusual in a way that significantly differentiates them from the hits of today.
Parsing through a set of songs with low HP/low popularity songs may uncover tracks that fall in the second category, well-produced but highly unusual and a strong candidate for sync licensing when TV shows and brands are seeking a very specific or distinctive approach.
This may be good for Music Supervisors on a budget, and also good for these artists that may otherwise not find a home for their music. Who knows? Perhaps placement in a TV show or commercial may help to launch or propel such an artist’s career and future royalty earnings.
High HP, Low Streams “Hidden Hits”
A song receiving a high HP score but low streams may indicate a track with great potential that has not yet been maximized – perhaps the artist’s marketing and/or branding strategy hasn’t been effective, or perhaps the release strategy was simply poorly timed or otherwise ineffective.
Such a song may be a candidate for a re-release, placement in a mainstream show, or even a re-recorded collaboration with a different artist in the same label to give it a fresh coat of paint.
This pilot study has interesting results that may provide the basis for a larger and more robust study examining the correlations between HP Score and various metrics of commercial success.
For now, it is mainly clear that songs with extremely high HP Score are streamed more and are generally written by more popular artists, while songs with extremely low HP Score tend to be streamed less and come from less popular artists.
Thinking creatively about what this means may be the key to record labels, song funds, and other large catalog owners fully maximizing the potential of their IP.
I am a composer and run a small distributed music, voice, and audio production team comprised of experts in 3 countries. I have a deep interest in startups, business, venture capital, networking, and learning about new ideas. I love meeting founders, angels, and people in VC and private equity to talk shop and connect people that need each other. I've Served clients and brands including Ubisoft, Garena (SEA), IGG Games, Hogarth Worldwide, We Are Social, Moving Bits, Playstudios, Ferrero, Samsung, GSK, Pernod Ricard, Pan Pacific Hotels etc. I’ve also worked with Wang Leehom, Joanna Dong, Derrick Hoh, Luke Slott, and a number of other artists. Founded and ran a professional orchestra in Boston for 5 years.